Problem-solving in mobile ad hoc networks
The ever-increasing complexity of network information systems has posed new highly demanding needs for better deployment of systems. This complexity is due not only to the increased number of interconnected users and devices, but also to the new forms of interactions among them. Under many circumstances, minor changes in this dynamic environment require manual intervention for the system's operation. Aiming at resolving this complexity explosion problem, the BISON project focused on developing self-organising and self-repairing robust network information systems. Mimicking the behaviour of insect communities and immune networks adaptation, self-organisation and robustness can be achieved without programming them into the individual artificial agents. Thereby, the global behaviour of large agent colonies may be adaptive and cooperative with arbitrary initial conditions, unforeseen scenarios, and variations in the environment or presence of deviant agents. One of the developed algorithms, the AntHocNet constitutes an example of this adaptive and robust behaviour with respect to network changes. Featuring a hybrid design, it involves a unique combination of reactivity and proactivity when anticipating and responding to sudden disruptive events in the most timely fashion. Based on adaptive learning mechanisms such as the Monte Carlo sampling and an information bootstrapping process it allows continuous adaptation of nodes' routing tables. Aided by a number of metrics including delay, throughput and signal-to-noise ratio, the AntHocNet offers efficient optimisation of the set-up of multiple routing paths. The innovative AntHocNet was tested and found to display superior performance under a number of open space and urban/structured scenarios of mobile ad hoc networks. Its hybrid and composite nature makes this algorithm suitable to use with various networks and network dynamics, as well as with several different modes/characteristics within the same heterogeneous network. For further information click at: http://www.cs.unibo.it/bison/
